Retail

AI Driven Automated Video Content Monitoring For Retail

VisionBot is a Cloud Based Articificial Intelligence (AI) and Computer Vision Platform that derives analysable objective data regarding objects, people and events from streaming or stored videos. Retail enterprises need continuous Visual content monitoring custom object and event detection through 24×7 visual monitoring to improve operational efficiency.

VisionBot leverages esiting Retail CCTV networks and the power of Deep Learning to achieve accuracy of detection and minimising false positives. We have built this platform based on our team’s deep skill sets and experience in data science, machine learning and cloud services. We remain in a continuous mode of bringing innovative new solutions around Machine Vision that helps improve operational efficiency of Enterprises.

VisionBot can also be offered as an on premise service in Hybrid / Private Cloud / Data Centre for enterprise requirements. It features an adaptive platform and learns on its own training. Users can also expedite training by providing preferences that makes it more efficient. works in both online (real time camera feed) or offline (stored video) modes. It offers the following features:

Four different use cases for retail CCTV based automated visual monitoring can be automated using camera feeds of the following sections, I.e. Entrance Sections, Billing Sections, Display Racks and Warehouses.

Visitors Information From Existing Entrance Gate Cameras:

From the CCTV feeds of the Entrance Gate cameras, VisionBot will help analyze and count the number of visitors entering over a specified period of time. We take into account the count of individuals coming in, what would be their possible age groups and also whether they are coming in groups. VisionBot Thus can provide the footfalls analytics of visitors to the mall. With this output we can help towards creating a customer friendly user experience and also help in providing data points for improving business development, non-intrusively.

Event Analysis From Existing Billing Section (cash Counter) Cameras:

From the CCTV feeds of the Billing Section cameras, VisionBot will help analyze movements taking place across the counter. This will provide an indication of potential financial transactions along with time stamp that had occurred at the Billing / Cash Counter.

The data derived from this analysis may be used during an internal audit or accounting reconciliation. The data may also be used to infer statistics and analytics of the Billing Section transactions and operations, non-intrusively.

Display Racks And Shopping Area Cameras

From the CCTV feeds of the Display Area cameras, VisionBot will help take a count of the number of people visiting a particular showcase over a period of time in a normalized manner. The gender and the age group of these people can also be detected. Analytics over a period of time will throw up trends in the shopping behaviours of customers. Thus this will help in displaying trending products and also in stocking sufficient quantity.

Warehouse Cameras:

CCTVs at warehouses need continuous monitoring and depending on that certain actions would need to be taken. Studies from several research institutions have concluded that personnel can only watch one monitor for 20 minutes, before losing focus. Also a human being can only monitor a maximum of 4 video streams simultaneously. Above that number, the chances of error in monitoring and event tracking, increases manifold. This can be automated using the VisionBot platform, 24×7 and for any number of cameras.

IMPLEMENTATION

VisionBot can be implemented in two ways

On Premise Edge Computing

We recommend the On Premise Data Center mode for installations that have or likely to add a large number of cameras. Cameras are easily installed either through our partners or by third party service providers. VisionBot works with most modern IP CCTV camera systems subject to adequate ambient light conditions.

VisionBot server(s) will be installed at local Data Center over a network either cabled or over wide area wifi.

Cloud Based Implementation

For smaller sites or temporary sites we recommend an asset light implementation of VisionBot. In this case there is no need of installing servers and data centre operations. Cameras can be directly interfaced to VisionBot cloud servers through appropriate internet routers. There is no upfront investments except cameras and switches. This configuration is highly popular and effective as it can also connect multiple sites that provides senior executive 24X7 access to field data and visuals.

VisionBot is a scalable system that can be used to analyze hundreds of camera feeds. It is cloud based thereby ensuring that Enterprise staff can access the data and information anywhere and anytime. Simultaneous monitoring of hundreds of cameras across multiple locations also provides deep insights for strategic planning, resource management and improves operation efficiency of organisations.